On Mon, Feb 02, 2026 at 02:50:28AM -0500, Michael S. Tsirkin wrote:
> On Sun, Feb 01, 2026 at 05:24:03PM -0800, Pierrick Bouvier wrote:
> > On 1/31/26 9:48 AM, Michael S. Tsirkin wrote:
> > > On Fri, Jan 30, 2026 at 06:00:56PM -0800, Pierrick Bouvier wrote:
> > > > Signed-off-by: Pierrick Bouvier <[email protected]>
> > > 
> > > Performance impact?
> > > 
> > > the reason we have these is for performance ...
> > > 
> > 
> > I would be very happy to run any benchmark that you might judge critical.
> > 
> > Should we run a disk read/write sequence with a virtio disk, or a download
> > with a virtio interface? Any other idea?
> 
> block for sure, people who care about network perf go the vhost or
> vhost user path.
> 
> So I CC'd Stefan Hajnoczi. Stefan do you feel this needs a test and what
> kind of test do you suggest as the most representative of I/O overhead?

This command-line lets you benchmark virtio-blk without actual I/O
slowing down the request processing:

  qemu-system-x86_64 \
      -M accel=kvm \
      -cpu host \
      -m 4G \
      --blockdev 
file,node-name=drive0,filename=boot.img,cache.direct=on,aio=native \
      --blockdev null-co,node-name=drive1,size=$((10 * 1024 * 1024 * 1024)) \
      --object iothread,id=iothread0 \
      --device virtio-blk-pci,drive=drive0,iothread=iothread0 \
      --device virtio-blk-pci,drive=drive1,iothread=iothread0

Here is a fio command-line for 4 KiB random reads:

  fio \
      --ioengine=libaio \
      --direct=1 \
      --runtime=30 \
      --ramp_time=10 \
      --rw=randread \
      --bs=4k \
      --iodepth=128 \
      --filename=/dev/vdb \
      --name=randread

This is just a single vCPU, but it should be enough to see if there is
any difference in I/O Operations Per Second (IOPS) or efficiency
(IOPS/CPU utilization).

Stefan

Attachment: signature.asc
Description: PGP signature

Reply via email to